Purpose: Two SCFSkp2 ubiquitin ligase–related proteins, Skp2 and cyclin-dependent kinase subunit 1 (Cks1), are involved in posttranscriptional degradation of p27Kip1 tumor suppressor. We analyzed the prognostic utility of p27Kip1 and its interacting cell cycle regulators in myxofibrosarcomas.

Experimental Design: Clinicopathologic features and tissue microarray–based immunohistochemical expression of p27Kip1, Skp2, Cks1, cyclin E, cyclin A, Ki-67, and minichromosome maintenance protein 2 (Mcm2) were assessed in 70 primary myxofibrosarcomas and correlated with clinical outcomes. Skp2 mRNA expression and the relationship between Skp2 and p27Kip1 proteins were examined in six cases by semiquantitative reverse transcription-PCR and Western blotting, respectively.

Results: High indices of Skp2 (≥10%), cyclin A (≥10%), and Mcm2 (≥50%) were adverse prognosticators at the univariate level. Furthermore, co-overexpression of Skp2 and cyclin A identified highly lethal cases in the entire cohort [P < 0.0001 for disease-specific survival (DSS), P = 0.0004 for overall survival (OS)] and the lower-grade subset (Fédération Nationale des Centres de Lutte Contre le Cancer grade 1 and 2; P = 0.0006 for DSS, P = 0.0093 for OS). In multivariate analyses, Skp2 overexpression overshadowed most intrinsic clinicopathologic factors and independently correlated with worse metastasis-free survival (P = 0.0012), DSS (P = 0.0234), and OS (P = 0.0056). Notably, positive margins independently predicted inferior local recurrence-free survival (P = 0.0012) and also negatively affected metastasis-free survival (P = 0.0471), DSS (P = 0.0152), and OS (P = 0.0173). Reverse transcription-PCR showed up-regulation of Skp2 mRNA in four cases and Western blotting displayed a matched expression pattern of Skp2.

Conclusions: Margin status and intrinsic property of myxofibrosarcomas both affect patient survival. Skp2 overexpression is highly representative of the biological aggressiveness of myxofibrosarcomas and plays an important prognostic role.

Myxofibrosarcoma is a common soft-tissue sarcoma usually affecting the extremities of the elderly (1, 2). Histologically, it shows a broad spectrum of nuclear atypia, cellularity, and mitoses and is characterized clinically by increased metastases after relentless local recurrences (1, 2). Limited series analyzing primary myxofibrosarcomas provided inconsistent prognostic information about various clinicopathologic factors (14). For instance, it is challenging to precisely predict patient survival by histologic grading for myxofibrosarcomas without multivariate survival analyses. Huang et al. (1) had compared the utility of two-, three- and four-tier grading schemes for myxofibrosarcomas at the low-grade end. However, their prognostic values were all found insignificant as compared with other variables (1). The limitations of histologic evaluation underscore the need to identify molecular markers that can effectively predict clinical outcomes of this entity.

During multistep carcinogenesis, sequential deregulation of multiple cell cycle regulators represents common modes of genetic alterations that promote progression in aggressive tumor subsets, thereby conferring an adverse effect on patient survival (5). p27Kip1 is a new tumor suppressor that specifically inhibits the activity of both cyclin E/CDK2 complex in G1-S transition and cyclin A/CDK2 complex in S phase to control cell cycle progression (6). Although reduced expression of p27Kip1 protein was shown to be highly associated with clinical aggressiveness in a variety of cancers (69), down-regulation of p27Kip1 mRNA was rarely observed in human malignancies (6). Instead, it becomes apparent that the loss of expression of p27Kip1 protein in malignant diseases resulted from enhanced ubiquitin-mediated proteolysis regulated by SCF-type ubiquitin E3 ligase complex (7, 8, 1014). The substrate-recognition component of the latter complex is Skp2, a member of the F-box protein family, which peaks in abundance during S phase to target Thr187-phosphorylated p27Kip1 for degradation (10, 11, 15). Recently, high Skp2 expression was also found predictive of poor prognosis in breast (12), prostate (13), kidney (7, 8), and head and neck cancers (10). Nevertheless, Skp2 did not correlate with low p27Kip1 in some malignancies (14, 16), implying that additional factors were implicated in p27Kip1 degradation or, alternatively, that Skp2 could possibly execute oncogenic function through p27Kip1-independent mechanisms. Recently, human cyclin-dependent kinase subunit 1 (Cks1) has been identified as a cofactor in the ubiquitination and degradation of p27Kip1, which confers an allosteric change in Skp2 to increase the affinity of SCFSKP2 complex to phosphorylated p27Kip1 (1719). The prognostic role of Cks1 has been exemplified in colorectal (18) and gastric (19) carcinomas in pilot studies, showing the association between Cks1 overexpression and adverse outcomes. To date, very few studies have addressed the prognostic utility of Skp2 or Cks1 in specific types of soft tissue sarcomas (14, 20).

The proliferation of tumor cells is highly related to the rate of DNA synthesis and may provide objective prognostic information. The minichromosome maintenance proteins (Mcm), consisting of six members (Mcm2-7), is a family of highly conserved proteins, which form a hexameric complex for regulating eukaryotic DNA duplication (21). Their expression is seen during all phases of cell cycle and may represent the rate of cell proliferation (21). To assemble the prereplication complex, replication licensing factors, Cdt1 and Cdc6, are required in G1 phase to load Mcm proteins onto chromatin. This process is coined as “licensing,” which ensures that DNA replicates only once during each cell cycle (2123). Interestingly, apart from p27Kip1, SCFSKP2 complex was also found to polyubiquitinate Cdt1 protein for proteasome-mediated degradation (24). Accordingly, it is tempting to apply immunohistochemistry to explore how p27Kip1/Skp2 pathway is related to the members of Mcm proteins in human cancers, including sarcomas.

Interestingly, a recent comparative genomic hybridization study on myxofibrosarcomas (25) reported high-level amplifications in chromosome 5p where the Skp2 gene is located. Although the prognostic values of p27Kip1and Skp2 had been individually reported for heterogeneous types of sarcomas (9, 14), no study has comprehensively elucidated the roles of ubiquitin-mediated proteolysis of p27Kip1 and its interacting proteins involving cell cycle regulation in myxofibrosarcomas. Therefore, the aims of this study were to specifically address the following questions about myxofibrosarcomas: (a) the expression status of p27Kip1, Skp2, Cks1, cyclin A, and cyclin E and two markers of cell proliferation, Ki-67 and Mcm2; and (b) the relative prognostic importance of clinicopathologic characteristics versus protein markers tested. Finally, we also examined Skp2 mRNA expression level by semiquantitative reverse transcription-PCR (RT-PCR) and the specificity and interaction of Skp2 and p27Kip1 proteins by Western blotting in six primary myxofibrosarcomas.

Patients and tumor material. Seventy patients with myxofibrosarcomas were retrospectively identified after pathologic review. They received primary surgical excision between 1985 and 2003 at three medical centers in Southern Taiwan, including Chang Gung Memorial Hospital (n = 42), Chi Mei Medical Center (n = 16), and Veteran General Hospital (n = 12). The diagnostic criteria were according to the latest WHO classification adapted from Mentzel's series (2, 26), requiring at least 10% of myxoid component featuring curvilinear vessels to designate myxofibrosarcomas. H&E-stained slides and paraffin-embedded tissue blocks of primary lesions were available in all cases. Institutional review boards of the involved hospitals approved retrospective clinical data collection and procurement of archival and fresh tissues. None of the patients had received neoadjuvant radiation or chemotherapy before surgery. Postoperative adjuvant radiotherapy (n = 26) and/or chemotherapy (n = 14) was randomly given in a minor subset of high-grade or recurrent cases without a consistent guideline, which were therefore excluded from analysis. The histopathologic features evaluated included the margin status, depth, percentages of myxoid area and spontaneous necrosis, nuclear pleomorphism (slight/moderate versus prominent), and mitoses per 10 high-power fields (HPF). In addition, grading and staging of primary myxofibrosarcomas were determined according to the updated Fédération Nationale des Centres de Lutte Contre le Cancer (FNCLCC) scheme (27) and the American Joint Committee on Cancer (AJCC) system (28), respectively.

Clinical follow-up data were available in 61 patients, which have been collected from the date of primary surgery until September 2004 or until the patients' death. The dates of each local recurrence and first metastasis were also recorded. The observation interval ranged from 2 to 201 months with a median of 32 months.

For semiquantitative RT-PCR and Western blot assays, six primary myxofibrosarcomas with paired normal and tumor tissue specimens were snap-frozen and stored at −80°C before RNA isolation and protein extraction.

Construction of tissue microarray blocks. To construct tissue microarray blocks, we used a recut H&E-stained slide from each block to define representative myxoid and solid regions. To circumvent tissue heterogeneity and to increase interpretable cases, six tissue cylinders (0.6 mm in diameter) for each specimen were punched from selected areas using a precision instrument (Beecher Instruments, Silver Spring, MD) and arrayed into two recipient blocks. In addition, we also arrayed three each of adenocarcinomas of the colorectum, lung, and gallbladder for the purposes of orientation and external positive controls in each tissue microarray block. These carcinomas were previously stained positive for the antibodies applied below. For negative controls, eight myxomas, including two cellular variants, were also arrayed together.

Immunohistochemistry. Sections of tissue microarray blocks were cut into an adhesive-coated glass slide system (instrumedicus, Hackensack, NJ) at 3-μm thickness. The slides were dewaxed in xylene and rehydrated through graded alcohols to water. For antigen retrieval, slides were pressure cooked in 10 mmol/L citrate buffer (pH 6) for 7 minutes. The slides were then washed using TBS buffer with 0.1% Tween 80 for this and subsequent washes. Endogenous peroxidase activity was quenched by 3% H2O2 treatment. After washing, the slides were incubated for 1 hour at room temperature with primary antibodies targeting Skp2 (2C8D9, 1:100; Zymed, San Francisco, CA), Cks1 (4G12G7, 1:250; Zymed), p27kip1 (1B4, 1:20; Novocastra, Newcastle, United Kingdom), cyclin A (6E6, 1:50; Novocastra), cyclin E (13A3, 1:40; Novocastra), Mcm2 (CRCT2.1, 1:25; Novocastra), and Ki-67 (MIB-1, 1:100; DAKO, Carpinteria, CA). Primary antibodies were detected using the ChemMate DAKO EnVision kit (DAKO, K5001). The slides were incubated with the secondary antibody for 30 minutes and developed with 3,3-diaminobenzidine for 5 minutes. Slides were then counterstained with Gill's hematoxylin. Coverslips were applied with CureMount mounting medium (instrumedicus). Incubation without the primary antibody was used as a negative control.

Assessment of immunohistochemical staining. One pathologist (C.F.L.), blinded to clinicopathologic data and patient outcomes, independently evaluated the tissue microarray slides. All cells of each tissue cylinder included in the tissue microarrays were evaluated and the percentage of tumor cells with definite moderate to intense nuclear immunoreactivity was recorded. Only cases containing two or more preserved tissue cores were scored, and scores from multiple cores for each marker in the same patient were averaged to obtain a mean labeling index (LI). By testing a series of different values (see Statistical analyses), the cutoffs of mean LIs to define overexpression or down-regulation of proteins were determined as follows: (a) Skp2 overexpression if ≥10% of tumor nuclei stained; (b) down-regulation of p27kip1 if <50% of tumor nuclei stained; (c) Cks1 overexpression if ≥20% of tumor nuclei stained; (d) cyclin A overexpression if ≥10% of tumor nuclei stained; (e) cyclin E overexpression if ≥10% of tumor nuclei stained; (f) high Ki-67 index if ≥30% of tumor nuclei stained; and (g) high Mcm2 index if ≥50% of tumor nuclei stained.

Semiquantitative RT-PCR. Total RNA was extracted by RNeasy Mini-kit (Qiagen, Valencia, CA) following the instruction of the manufacturer and the amount of mRNA was measured by UV spectrophotometer. Two micrograms of total RNA were subjected to RT-PCR using Qiagen one-step RT-PCR kit (Qiagen). The sequences of Skp2 forward and reverse primers were 5′-CCTATCACTCAGTCGGTGCTATGA-3′ and 5′-GGGTACCATCTGGCACGATT-3′, respectively. The RT-PCR conditions included 50°C for 30 minutes for reverse transcription; 95°C for 15 minutes; 28 cycles of 95°C for 45 seconds, 57°C for 45 seconds, and 72°C for 1 minute; and 72°C for 7 minutes. The mRNA of porphobilinogen deaminase (PBGD) gene was in parallel reverse transcribed and amplified as an internal control using the forward primer 5′-TGTCTGGTAACGGCAATGCGGCTGCAAC-3′ and the reverse primer 5′-TCAATGTTGCCACCACACTGTCCGTCT-3′. The expected sizes of the amplified transcripts were 109 bp for Skp2 and 127 bp for PBGD. The PCR products were separated on 2.0% agarose gels and the density of each product was measured by densitometry (Bio-Rad, Shelton, CT). The Skp2 mRNA expression ratio in each myxofibrosarcoma was defined as (TSkp2 / TPBGD) / (NSkp2 / NPBGD), and Skp2 transcript was regarded as up-regulated when the ratio was ≥3.

Western blot assays. Equal amounts of protein extract were separated on SDS-8% PAGE, transferred to nitrocellulose membranes, and then blocked with 5% skimmed milk in TBS-0.1% Tween 80 buffer at room temperature for 1 hour. Afterwards, the membranes were probed with antibodies against Skp2 (1:250, H-435; Santa Cruz Biotechnology, Santa Cruz, CA), p27Kip1(1:500, 1 B4; Novocastra), and housekeeping glyceraldehyde-3-phosphate dehydrogenase proteins (MAB347, 1:3,000; Chemicon, Temecula, CA) at 4°C overnight and then incubated with the secondary antibody at room temperature for 1 hour. Enhanced chemiluminescence reagents (Amersham, Piscataway, NJ) were used to visualize the targeted proteins, which were then quantitated by densitometry.

Statistical analyses. Statistical analyses were done using the SPSS 10 software package. Associations among various variables were assessed using Pearson correlation coefficient test, Mann-Whitney U test, or Kruskal-Wallis method as appropriate. For survival analyses, the end points analyzed were local recurrence-free survival (LRFS), metastasis-free survival (MeFS), disease-specific survival (DSS), and overall survival (OS). A series of cutoff values were tested for continuous variables, such as mitotic activity and mean LIs of markers. The cutoffs giving the best P values were adopted to construct Kaplan-Meier curves and compare prognostic differences by univariate log-rank test. All significant variables at univariate level were entered into Cox multivariate regression model to analyze their relative risk and independent prognostic effect in a stepwise forward fashion. For all analyses, two-sided tests of significance were used with P < 0.05 considered as significant.

Clinicopathologic features. The cohort consisted of 38 males and 32 females with the median age being 64 years (range, 16-83 years). The primary tumors were located in the extremities in 58 cases (83%) and axial in 12 (17%) with the lower limb being the most common site (n = 41; 59%). Thirty tumors (43%) were superficial and 40 (57%) were deep seated. The size of primary tumors was available in 64 cases and ranged from 1.2 to 24 cm (median, 5.5 cm); 29 cases (45.3%) had tumors measuring ≥6 cm in the greatest dimension. For the remaining six cases, one was grossly described as a large, multinodular mass with the predominant nodule being 8.5 cm whereas the other five were of unknown size.

Histologically, the tumors were composed of sarcomatous cells with a wide spectrum of nuclear pleomorphism set in a variably myxoid matrix containing long curvilinear vessels (Fig. 1A-C). The percentage of myxoid area in each primary tumor ranged from 10% to 100% (median, 50%) and 16 cases containing ≥75% of myxoid component were considered as predominantly myxoid. The majority of cases (n = 54, 77%) displayed tumor cells with prominent nuclear pleomorphism as evidenced by the presence of bizarre giant cells regardless of the number. Remarkable tumor necrosis, defined as ≥10% of surface areas, was observed in 23 cases (33%); 47 cases showed either minimal (<10%, n = 7) or no tumor necrosis (n = 40). Mitotic counts revealed a variation from 0 to 50 (median, 9) per 10 HPFs and were 0 to 9 per 10 HPFs in 36 cases, 10 to 19 per 10 HPFs in 22 cases, and ≥20 per 10 HPFs in 12 cases. The surgical procedures consisted of wide excision or amputation with clear margins in 43 patients and marginal or intralesional excision with positive microscopic margins in 26 whereas the status of margin could not be ascertained in the remaining one. By FNCLCC grading, primary tumors were classified as grade 1 in 25 cases (35.7%; Fig. 1A), grade 2 (51.4%; Fig. 1B) in 36 cases, and grade 3 (12.9%; Fig. 1C) in 9 cases. AJCC stage could be assessed in 65 cases, 9 (13.8%), 27 (41.5%), and 29 (44.7%) of which were stage I, stage II, and stage III, respectively. There was no patient with metastatic disease at initial diagnosis.

Fig. 1.

H&E (×100) and immunohistochemical (×250) stains of representative myxofibrosarcomas, ranging from FNCLCC grade 1 to grade 3. A, grade 1 tumor showing low cellularity, mild nuclear atypia, and abundant myxoid matrix. B, grade 2 tumor featuring increased cellularity and moderate nuclear atypia with characteristic long curvilinear vessels. C, grade 3 tumor displaying large pleomorphic tumor giant cells with increased mitoses and solid areas. Weak expressions of Skp2, Cks1, and cyclin A are noted in a grade 1 tumor (D, J, and M, respectively) whereas grade 2 (E, K, and N, respectively) and grade 3 (F, L, and O, respectively) lesions show overexpressions of Skp2, Cks1, and cyclin A. p27Kip1 expression is preserved in grade 1 (G) and grade 2 (H) tumors but lost in a grade 3 lesion (I). Mcm2 stains showed a gradual increase in LI in grade 1 (P), grade 2 (Q), and grade 3 (R) myxofibrosarcomas.

Fig. 1.

H&E (×100) and immunohistochemical (×250) stains of representative myxofibrosarcomas, ranging from FNCLCC grade 1 to grade 3. A, grade 1 tumor showing low cellularity, mild nuclear atypia, and abundant myxoid matrix. B, grade 2 tumor featuring increased cellularity and moderate nuclear atypia with characteristic long curvilinear vessels. C, grade 3 tumor displaying large pleomorphic tumor giant cells with increased mitoses and solid areas. Weak expressions of Skp2, Cks1, and cyclin A are noted in a grade 1 tumor (D, J, and M, respectively) whereas grade 2 (E, K, and N, respectively) and grade 3 (F, L, and O, respectively) lesions show overexpressions of Skp2, Cks1, and cyclin A. p27Kip1 expression is preserved in grade 1 (G) and grade 2 (H) tumors but lost in a grade 3 lesion (I). Mcm2 stains showed a gradual increase in LI in grade 1 (P), grade 2 (Q), and grade 3 (R) myxofibrosarcomas.

Close modal

Profiling of Immunohistochemical expression and associations with clinicopathologic variables. Expressions of Skp2, Cks1, cyclin A, and cyclin E were found only in myxofibrosarcomas but not in any myxoma. However, nuclear staining for p27Kip1 protein could be sporadically detected in tumor cells of myxomas. The LI of Skp2 ranged from 0% to 85% of the sarcoma cells (median, 11%; Fig. 1D-F) and distinct overexpression of Skp2 was identified in 34 of 67 cases (50.7%; Fig. 1E and F) interpretable for scoring. The expression of p27Kip1 displayed a wide variation in LI from 0% to 100% (median, 50%; Fig. 1G-I) and down-regulation of p27Kip1 was identified in 33 of 70 cases (47.1%; Fig. 1I). For cases suitable for scoring, overexpressions of Cks1 (median LI, 17.5%; Fig. 1J-L), cyclin A (median LI, 11%; Fig. 1M-O), cyclin E (median LI, 4.5%), Ki-67 (median LI, 18.5%), and Mcm2 (median LI, 35%; Fig. 1P-R) were observed in 34 of 70 (48.5%), 40 of 69 (59.7%), 24 of 70 (34.3%), 24 of 68 (35.2%), and 19 of 69 (27.5%) primary tumors, respectively.

The results of immunohistochemical staining and associations with clinicopathologic variables are listed in Table 1. The increment of FNCLCC grade was found positively related to the expression levels of Skp2 (Fig. 2A), Cks1, cyclin A, cyclin E, and Mcm2 but not associated with the expression of p27Kip1 or Ki-67. Noticeably, the former five significant markers were also robust in distinguishing between grade 1 and grade 2 cases. To a lesser degree, the expression levels of Skp2, Cks1, and Mcm2 also increased significantly with AJCC stage. In addition, significant associations were observed between prominent nuclear pleomorphism and the high LIs of Skp2, cyclin A, cyclin E, Ki-67, and Mcm2. On the other hand, tumor size ≥6 cm and deep-seated tumors were merely associated with high cyclin E and high Skp2 expression, respectively. The LIs of Skp2, Cks1, cyclin A, and Mcm2 revealed significant positive correlations with both the percentages of necrotic area and mitotic count. However, the LIs of cyclin E and Ki-67 were only positively associated with tumor necrosis and mitotic activity, respectively. In contrast, the percentage of myxoid area was inversely associated with the LIs of Cks1, cyclin A, cyclin E, Ki-67, and Mcm2. However, there was no significant correlation between the markers investigated and sex, age, or the microscopic status of surgical margins. These findings supported that the aforementioned cell cycle regulators and proliferative markers in general were significantly indicative of tumor progression, implying a substantial role of multiple genetic aberrations in tumorigenesis of myxofibrosarcomas.

Table 1.

Associations between immunohistochemical markers and clinicopathologic factors

Skp2p27Kip1Cks1Cyclin ACyclin EKi-67Mcm2
Sex* P = 0.935 P = 0.595 P = 0.571 P = 0.526 P = 0.924 P = 0.458 P = 0.396 
    Male 15.00 ± 16.24 (n = 37) 45.00 ± 25.69 (n = 38) 25.18 ± 19.84 (n = 38) 12.61 ± 10.07 (n = 38) 9.63 ± 14.17 (n = 38) 24.11 ± 16.48 (n = 38) 32.95 ± 24.32 (n = 37) 
    Female 15.00 ± 14.61 (n = 30) 47.72 ± 22.57 (n = 32) 25.66 ± 27.17 (n = 32) 14.45 ± 10.91 (n = 31) 6.44 ± 7.62 (n = 32) 21.60 ± 17.18 (n = 30) 39.34 ± 27.59 (n = 32) 
Surgical margins* P = 0.302 P = 0.738 P = 0.142 P = 0.677 P = 0.841 P = 0.403 P = 0.401 
    Negative 15.17 ± 12.31 (n = 41) 46.58 ± 23.98 (n = 43) 22.98 ± 23.82 (n = 43) 13.31 ± 10.70 (n = 42) 7.93 ± 11.68 (n = 43) 24.31 ± 17.17 (n = 42) 33.33 ± 25.37 (n = 43) 
    Positive 15.16 ± 19.88 (n = 25) 45.54 ± 25.41 (n = 26) 29.08 ± 22.77 (n = 26) 13.88 ± 10.27 (n = 26) 8.65 ± 12.10 (n = 26) 20.06 ± 16.31 (n = 25) 39.32 ± 26.84 (n = 25) 
Tumor size (cm)* P = 0.111 P = 0.299 P = 0.182 P = 0.362 P = 0.0307 P = 0.698 P = 0.157 
    <6 11.45 ± 10.90 (n = 33) 49.14 ± 25.06 (n = 35) 23.43 ± 23.73 (n = 35) 11.91 ± 8.92 (n = 34) 6.86 ± 11.91(n = 35) 23.61 ± 16.65 (n = 33) 33.29 ± 26.24 (n = 34) 
    ≥6 18.36 ± 18.18 (n = 28) 44.17 ± 23.47 (n = 29) 29.97 ± 23.11 (n = 29) 15.45 ± 11.95 (n = 29) 10.55 ± 12.09 (n = 29) 25.24 ± 16.97 (n = 29) 41.03 ± 24.80 (n = 29) 
Tumor depth* P = 0.023 P = 0.934 P = 0.660 P = 0.584 P = 0.339 P = 0.309 P = 0.068 
    Superficial 10.79 ± 12.66 (n = 28) 45.00 ± 27.35 (n = 30) 24.87 ± 24.38 (n = 30) 12.14 ± 8.68 (n = 29) 8.00 ± 12.95 (n = 30) 20.54 ± 15.91 (n = 28) 30.03 ± 27.48 (n = 30) 
    Deep 18.02 ± 16.62 (n = 39) 47.18 ± 21.81 (n = 40) 25.78 ± 22.76 (n = 40) 14.38 ± 11.53 (n = 40) 8.30 ± 10.79 (n = 40) 24.73 ± 17.23 (n = 40) 40.44 ± 23.98 (n = 39) 
Nuclear pleomorphism* P = 0.017 P = 0.413 P = 0.060 P = 0.011 P = 0.005 P = 0.019 P = 0.043 
    Mild/moderate 7.75 ± 17.27 (n = 16) 41.50 ± 22.55 (n = 16) 17.13 ± 21.09 (n = 16) 7.67 ± 7.15 (n = 15) 3.06 ± 4.70 (n = 16) 14.43 ± 15.87 (n = 14) 24.56 ± 23.27 (n = 14) 
    Prominent 17.27 ± 16.47 (n = 51) 47.65 ± 24.66 (n = 54) 27.83 ± 23.54 (n = 54) 15.04 ± 10.66 (n = 54) 9.69 ± 12.69 (n = 54) 25.22 ± 16.33 (n = 54) 39.34 ± 25.86 (n = 53) 
FNCLCC grade P = 0.002 P = 0.863 P = 0.039 P < 0.001 P = 0.004 P = 0.125 P = 0.009 
    Grade 1 7.66 ± 7.42 (n = 24) 44.08 ± 26.06 (n = 25) 16.04 ± 15.60 (n = 25) 6.96 ± 4.94 (n = 24) 3.48 ± 6.02 (n = 25) 17.38 ± 14.47 (n = 24) 25.04 ± 22.62 (n = 25) 
    Grade 2 15.91 ± 13.21(n = 34) 47.75 ± 23.59 (n = 25) 29.42 ± 15.60 (n = 36) 15.31 ± 10.62 (n = 36) 10.53 ± 12.66 (n = 36) 25.26 ± 16.95 (n = 35) 38.74 ± 24.89 (n = 35) 
    Grade 3 31.11 ± 24.99 (n = 9) 46.22 ± 23.34 (n = 36) 35.22 ± 24.38 (n = 9) 23.22 ± 10.47 (n = 9) 11.78 ± 16.05 (n = 9) 29.22 ± 118.82 (n = 9) 55.11 ± 26.69 (n = 9) 
    Total 15.00 ± 15.41(n = 67) 46.24 ± 24.18 (n = 70) 25.39 ± 23.30 (n = 70) 13.43 ± 10.42 (n = 69) 8.17 ± 11.68 (n = 70) 23.00 ± 16.71 (n = 68) 35.91 ± 25.89 (n = 69) 
AJCC stage P = 0.022 P = 0.517 P = 0.044 P = 0.063 P = 0.097 P = 0.151 P = 0.018 
    Stage I 6.44 ± 7.50 (n = 9) 38.89 ± 21.94 (n = 9) 10.33 ± 11.03 (n = 9) 6.13 ± 4.45 (n = 8) 3.56 ± 4.59 (n = 9) 14.25 ± 15.07 (n = 8) 19.67 ± 21.77 (n = 9) 
    Stage II 11.68 ± 10.90 (n = 25) 48.48 ± 25.65 (n = 27) 27.78 ± 23.05 (n = 27) 11.81 ± 7.41 (n = 27) 8.33 ± 13.26 (n = 27) 23.73 ± 15.29 (n = 26) 33.61 ± 24.98 (n = 26) 
    Stage III 19.75 ± 17.97 (n = 28) 47.93 ± 23.74 (n = 29) 29.59 ± 25.01 (n = 29) 16.55 ± 12.45 (n = 29) 9.82 ± 12.10 (n = 29) 26.17 ± 16.91 (n = 29) 45.24 ± 24.30 (n = 29) 
    Total 14.56 ± 14.92 (n = 62) 46.91 ± 24.18 (n = 65) 26.17 ± 23.38 (n = 65) 13.25 ± 10.30 (n = 65) 8.34 ± 11.93 (n = 65) 23.65 ± 16.23 (n = 63) 36.92 ± 25.48 (n = 64) 
Age P = 0.554, r = −0.554 (n = 70) P = 0.183, r = 0.161 (n = 70) P = 0.330, r = 0.118 (n = 70) P = 0.415, r = −0.100 (n = 69) P = 0.131, r = −0.182 (n = 70) P = 0.991, r = 0.001 (n = 68) P = 0.851, r = −0.023 (n = 70) 
% of tumor necrosis P < 0.001, r = 0.422 (n = 67) P = 0.138, r = −0.179 (n = 70) P = 0.024, r = 0.270 (n = 70) P = 0.007, r = 0.322 (n = 69) P = 0.001, r = 0.378 (n = 70) P = 0.120, r = 0.191 (n = 68) P = 0.022, r = 0.275 (n = 69) 
Mitotic rate P < 0.001, r = 0.422 (n = 67) P = 0.949, r = 0.008 (n = 70) P = 0.008, r = 0.319 (n = 70) P < 0.001, r = 0.494 (n = 69) P = 0.360, r = 0.113 (n = 70) P = 0.019, r = 0.287 (n = 68) P = 0.004, r = 0.348 (n = 67) 
% of myxoid area P = 0.103, r = −0.201 (n = 67) P = 0.248, r = 0.140 (n = 70) P = 0.003, r = −0.354 (n = 70) P = 0.001, r = −0.408 (n = 69) P = 0.031, r = −0.258 (n = 70) P = 0.004, r = −0.343 (n = 70) P = 0.001, r = −0.385 (n = 69) 
Skp2p27Kip1Cks1Cyclin ACyclin EKi-67Mcm2
Sex* P = 0.935 P = 0.595 P = 0.571 P = 0.526 P = 0.924 P = 0.458 P = 0.396 
    Male 15.00 ± 16.24 (n = 37) 45.00 ± 25.69 (n = 38) 25.18 ± 19.84 (n = 38) 12.61 ± 10.07 (n = 38) 9.63 ± 14.17 (n = 38) 24.11 ± 16.48 (n = 38) 32.95 ± 24.32 (n = 37) 
    Female 15.00 ± 14.61 (n = 30) 47.72 ± 22.57 (n = 32) 25.66 ± 27.17 (n = 32) 14.45 ± 10.91 (n = 31) 6.44 ± 7.62 (n = 32) 21.60 ± 17.18 (n = 30) 39.34 ± 27.59 (n = 32) 
Surgical margins* P = 0.302 P = 0.738 P = 0.142 P = 0.677 P = 0.841 P = 0.403 P = 0.401 
    Negative 15.17 ± 12.31 (n = 41) 46.58 ± 23.98 (n = 43) 22.98 ± 23.82 (n = 43) 13.31 ± 10.70 (n = 42) 7.93 ± 11.68 (n = 43) 24.31 ± 17.17 (n = 42) 33.33 ± 25.37 (n = 43) 
    Positive 15.16 ± 19.88 (n = 25) 45.54 ± 25.41 (n = 26) 29.08 ± 22.77 (n = 26) 13.88 ± 10.27 (n = 26) 8.65 ± 12.10 (n = 26) 20.06 ± 16.31 (n = 25) 39.32 ± 26.84 (n = 25) 
Tumor size (cm)* P = 0.111 P = 0.299 P = 0.182 P = 0.362 P = 0.0307 P = 0.698 P = 0.157 
    <6 11.45 ± 10.90 (n = 33) 49.14 ± 25.06 (n = 35) 23.43 ± 23.73 (n = 35) 11.91 ± 8.92 (n = 34) 6.86 ± 11.91(n = 35) 23.61 ± 16.65 (n = 33) 33.29 ± 26.24 (n = 34) 
    ≥6 18.36 ± 18.18 (n = 28) 44.17 ± 23.47 (n = 29) 29.97 ± 23.11 (n = 29) 15.45 ± 11.95 (n = 29) 10.55 ± 12.09 (n = 29) 25.24 ± 16.97 (n = 29) 41.03 ± 24.80 (n = 29) 
Tumor depth* P = 0.023 P = 0.934 P = 0.660 P = 0.584 P = 0.339 P = 0.309 P = 0.068 
    Superficial 10.79 ± 12.66 (n = 28) 45.00 ± 27.35 (n = 30) 24.87 ± 24.38 (n = 30) 12.14 ± 8.68 (n = 29) 8.00 ± 12.95 (n = 30) 20.54 ± 15.91 (n = 28) 30.03 ± 27.48 (n = 30) 
    Deep 18.02 ± 16.62 (n = 39) 47.18 ± 21.81 (n = 40) 25.78 ± 22.76 (n = 40) 14.38 ± 11.53 (n = 40) 8.30 ± 10.79 (n = 40) 24.73 ± 17.23 (n = 40) 40.44 ± 23.98 (n = 39) 
Nuclear pleomorphism* P = 0.017 P = 0.413 P = 0.060 P = 0.011 P = 0.005 P = 0.019 P = 0.043 
    Mild/moderate 7.75 ± 17.27 (n = 16) 41.50 ± 22.55 (n = 16) 17.13 ± 21.09 (n = 16) 7.67 ± 7.15 (n = 15) 3.06 ± 4.70 (n = 16) 14.43 ± 15.87 (n = 14) 24.56 ± 23.27 (n = 14) 
    Prominent 17.27 ± 16.47 (n = 51) 47.65 ± 24.66 (n = 54) 27.83 ± 23.54 (n = 54) 15.04 ± 10.66 (n = 54) 9.69 ± 12.69 (n = 54) 25.22 ± 16.33 (n = 54) 39.34 ± 25.86 (n = 53) 
FNCLCC grade P = 0.002 P = 0.863 P = 0.039 P < 0.001 P = 0.004 P = 0.125 P = 0.009 
    Grade 1 7.66 ± 7.42 (n = 24) 44.08 ± 26.06 (n = 25) 16.04 ± 15.60 (n = 25) 6.96 ± 4.94 (n = 24) 3.48 ± 6.02 (n = 25) 17.38 ± 14.47 (n = 24) 25.04 ± 22.62 (n = 25) 
    Grade 2 15.91 ± 13.21(n = 34) 47.75 ± 23.59 (n = 25) 29.42 ± 15.60 (n = 36) 15.31 ± 10.62 (n = 36) 10.53 ± 12.66 (n = 36) 25.26 ± 16.95 (n = 35) 38.74 ± 24.89 (n = 35) 
    Grade 3 31.11 ± 24.99 (n = 9) 46.22 ± 23.34 (n = 36) 35.22 ± 24.38 (n = 9) 23.22 ± 10.47 (n = 9) 11.78 ± 16.05 (n = 9) 29.22 ± 118.82 (n = 9) 55.11 ± 26.69 (n = 9) 
    Total 15.00 ± 15.41(n = 67) 46.24 ± 24.18 (n = 70) 25.39 ± 23.30 (n = 70) 13.43 ± 10.42 (n = 69) 8.17 ± 11.68 (n = 70) 23.00 ± 16.71 (n = 68) 35.91 ± 25.89 (n = 69) 
AJCC stage P = 0.022 P = 0.517 P = 0.044 P = 0.063 P = 0.097 P = 0.151 P = 0.018 
    Stage I 6.44 ± 7.50 (n = 9) 38.89 ± 21.94 (n = 9) 10.33 ± 11.03 (n = 9) 6.13 ± 4.45 (n = 8) 3.56 ± 4.59 (n = 9) 14.25 ± 15.07 (n = 8) 19.67 ± 21.77 (n = 9) 
    Stage II 11.68 ± 10.90 (n = 25) 48.48 ± 25.65 (n = 27) 27.78 ± 23.05 (n = 27) 11.81 ± 7.41 (n = 27) 8.33 ± 13.26 (n = 27) 23.73 ± 15.29 (n = 26) 33.61 ± 24.98 (n = 26) 
    Stage III 19.75 ± 17.97 (n = 28) 47.93 ± 23.74 (n = 29) 29.59 ± 25.01 (n = 29) 16.55 ± 12.45 (n = 29) 9.82 ± 12.10 (n = 29) 26.17 ± 16.91 (n = 29) 45.24 ± 24.30 (n = 29) 
    Total 14.56 ± 14.92 (n = 62) 46.91 ± 24.18 (n = 65) 26.17 ± 23.38 (n = 65) 13.25 ± 10.30 (n = 65) 8.34 ± 11.93 (n = 65) 23.65 ± 16.23 (n = 63) 36.92 ± 25.48 (n = 64) 
Age P = 0.554, r = −0.554 (n = 70) P = 0.183, r = 0.161 (n = 70) P = 0.330, r = 0.118 (n = 70) P = 0.415, r = −0.100 (n = 69) P = 0.131, r = −0.182 (n = 70) P = 0.991, r = 0.001 (n = 68) P = 0.851, r = −0.023 (n = 70) 
% of tumor necrosis P < 0.001, r = 0.422 (n = 67) P = 0.138, r = −0.179 (n = 70) P = 0.024, r = 0.270 (n = 70) P = 0.007, r = 0.322 (n = 69) P = 0.001, r = 0.378 (n = 70) P = 0.120, r = 0.191 (n = 68) P = 0.022, r = 0.275 (n = 69) 
Mitotic rate P < 0.001, r = 0.422 (n = 67) P = 0.949, r = 0.008 (n = 70) P = 0.008, r = 0.319 (n = 70) P < 0.001, r = 0.494 (n = 69) P = 0.360, r = 0.113 (n = 70) P = 0.019, r = 0.287 (n = 68) P = 0.004, r = 0.348 (n = 67) 
% of myxoid area P = 0.103, r = −0.201 (n = 67) P = 0.248, r = 0.140 (n = 70) P = 0.003, r = −0.354 (n = 70) P = 0.001, r = −0.408 (n = 69) P = 0.031, r = −0.258 (n = 70) P = 0.004, r = −0.343 (n = 70) P = 0.001, r = −0.385 (n = 69) 

NOTE: LI of each marker was expressed as mean ± SD.

*

Mann-Whitney U test.

Kruskal-Wallis test.

Pearson correlation coefficient test.

Fig. 2.

A, Skp2 LI for each grade of myxofibrosarcoma showing a significant correlation (P = 0.002). B, Skp2 LI is positively correlated with the expression level of Mcm2 (P < 0.001, r = 0.351). C, a close correlation was observed between the LIs of Mcm2 and Ki-67 (P < 0.001, r = 0.705).

Fig. 2.

A, Skp2 LI for each grade of myxofibrosarcoma showing a significant correlation (P = 0.002). B, Skp2 LI is positively correlated with the expression level of Mcm2 (P < 0.001, r = 0.351). C, a close correlation was observed between the LIs of Mcm2 and Ki-67 (P < 0.001, r = 0.705).

Close modal

Correlations among immunohistochemical markers.Table 2 summarizes the correlations between two markers of cell proliferation, Ki-67 and Mcm2, and five cell cycle regulators. The LIs of Skp2 (Fig. 2B), Cks1, cyclin A, and cyclin E were all positively associated with both Ki-67 and Mcm2 expression. Additionally, the expression of Mcm2 was highly correlated with LI of Ki-67 (Fig. 2C) and it tended to express at higher levels and distribute over a broader range than Ki-67. Except for Ki-67, the LI of p27Kip1 was not inversely associated with Skp2 or with Cks1 expression.

Table 2.

Correlations between LIs of cell cycle regulators and of markers of cell proliferation

Ki-67Mcm2
Skp2 P = 0.004, r = 0.490 P < 0.001, r = 0.351 
p27Kip1 P = 0.024, r = −0.273 P = 0.055, r = −0.055 
Cks1 P < 0.001, r = 0.419 P < 0.001, r = 0.478 
Cyclin A P < 0.001, r = 0.552 P < 0.001, r = 0.618 
Cyclin E P = 0.003, r = 0.353 P < 0.001, r = 0.400 
Mcm2 P < 0.001, r = 0.705 — 
Ki-67Mcm2
Skp2 P = 0.004, r = 0.490 P < 0.001, r = 0.351 
p27Kip1 P = 0.024, r = −0.273 P = 0.055, r = −0.055 
Cks1 P < 0.001, r = 0.419 P < 0.001, r = 0.478 
Cyclin A P < 0.001, r = 0.552 P < 0.001, r = 0.618 
Cyclin E P = 0.003, r = 0.353 P < 0.001, r = 0.400 
Mcm2 P < 0.001, r = 0.705 — 

NOTE: All comparisons were examined by Pearson correlation coefficient test.

Univariate survival analyses. Among the 61 patients with follow-up, 27 patients developed at least one local recurrence and 14 patients experienced distant metastasis. The sites of first metastasis were lung in eight patients and lymph nodes, somatic soft tissue, and abdominal cavity in two patients each. At last follow-up, 13 patients died of myxofibrosarcomas whereas 4 patients died of other causes. The cumulative 3-year rates of LRFS, MeFS, DSS, and OS were 51%, 72%, 80%, and 74%, respectively. Correlations of clinical, histologic, and immunohistochemical prognostic factors to various end point survivals are shown in Table 3 and Fig. 3A to H. By virtue of the small number of FNCLCC grade 3 cases, they were grouped with grade 2 cases (Fig. 3C) for subsequent analyses with no given statistical difference in patient survival between grade 1 and grade 2 (Fig. 3B).

Table 3.

Results of univariate log-rank analysis of prognostic factors in patients with follow-up data

FactorsNo. patientsLRFS
MeFS
DSS
OS
No. eventsPNo. eventsPNo. eventsPNo. eventsP
Sex          
    Male 29 14 0.4994 12 0.5592 0.8208 0.9532 
    Female 32 16  11    
Age (y)          
    <60 25 10 0.1588 0.1412 0.9706 0.3955 
    ≥60 36 20  16   11  
Location          
    Extremity 53 24 0.0899 20 0.9742 10 0.2672 14 0.6097 
    Central     
Tumor size (cm)          
    <6 30 16 0.3475 0.0143 0.1951 0.5226 
    ≥6 25 110  12    
Tumor depth          
    Superficial 26 15 0.5416 0.1388 0.3505 0.1384 
    Deep 35 15  15   11  
Margin          
    Clear 39 14 0.0003 10 0.0295 0.0438 0.0084 
    Positive 21 16  13   11  
Nuclear pleomorphism          
    Mild/moderate 14 0.3553 0.0129 0.1228 0.0412 
    Prominent 47 25  22  12  16  
Myxoid area          
    ≥75% 13 0.0653 0.0172 0.0997 0.1423 
    <75% 48 25  20  12  15  
Necrotic area          
    <10% 41 23 0.4081 12 0.0005 0.0324 11 0.1131 
    ≥10% 20  11    
Mitotic counts          
    <20/10 HPFs 50 24 0.1140 16 0.0025 0.0006 12 0.0060 
    ≥20/10 HPFs 11     
FNCLCC grade          
    Grade 1 21 10 0.5220 0.0078 0.0174 0.0200 
    Grade 2 and 3 40 20  20  12  15  
AJCC stage          
    Stage 1 and 2 32 18 0.2800 0.0470 0.2769 0.1730 
    Stage 3 25  11    
Skp2          
    <10% 29 11 0.0585 <0.0001 0.0010 0.0020 
    ≥10% 30 17  18  11  12  
Cyclin A          
    <10% 25 0.0320 0.0815 0.0075 0.0353 
    ≥10% 35 21  17  12  14  
Skp2/cyclin A          
    Neither or either ≥10% 35     <0.0001 0.0004 
    Both≥10% 23     11  11  
Cyclin E          
    <10% 38 19 0.8624 13 0.3306 0.6670 11 0.7979 
    ≥10% 23 11  10    
Cks1          
    <20% 31 14 0.3443 12 0.8295 0.3281 10 0.8766 
    ≥20% 30 16  11    
p27Kip1          
    <50% 29 16 0.5155 12 0.5642 87 0.6362 0.5528 
    ≥50% 32 14  11    
Ki-67          
    <30% 39 20 0.6634 14 0.0781 0.5324 13 0.8866 
    ≥30% 20     
Mcm2          
    <50% 43 22 0.8333 13 0.0304 10 0.8771 11 0.3313 
    ≥50% 17     
FactorsNo. patientsLRFS
MeFS
DSS
OS
No. eventsPNo. eventsPNo. eventsPNo. eventsP
Sex          
    Male 29 14 0.4994 12 0.5592 0.8208 0.9532 
    Female 32 16  11    
Age (y)          
    <60 25 10 0.1588 0.1412 0.9706 0.3955 
    ≥60 36 20  16   11  
Location          
    Extremity 53 24 0.0899 20 0.9742 10 0.2672 14 0.6097 
    Central     
Tumor size (cm)          
    <6 30 16 0.3475 0.0143 0.1951 0.5226 
    ≥6 25 110  12    
Tumor depth          
    Superficial 26 15 0.5416 0.1388 0.3505 0.1384 
    Deep 35 15  15   11  
Margin          
    Clear 39 14 0.0003 10 0.0295 0.0438 0.0084 
    Positive 21 16  13   11  
Nuclear pleomorphism          
    Mild/moderate 14 0.3553 0.0129 0.1228 0.0412 
    Prominent 47 25  22  12  16  
Myxoid area          
    ≥75% 13 0.0653 0.0172 0.0997 0.1423 
    <75% 48 25  20  12  15  
Necrotic area          
    <10% 41 23 0.4081 12 0.0005 0.0324 11 0.1131 
    ≥10% 20  11    
Mitotic counts          
    <20/10 HPFs 50 24 0.1140 16 0.0025 0.0006 12 0.0060 
    ≥20/10 HPFs 11     
FNCLCC grade          
    Grade 1 21 10 0.5220 0.0078 0.0174 0.0200 
    Grade 2 and 3 40 20  20  12  15  
AJCC stage          
    Stage 1 and 2 32 18 0.2800 0.0470 0.2769 0.1730 
    Stage 3 25  11    
Skp2          
    <10% 29 11 0.0585 <0.0001 0.0010 0.0020 
    ≥10% 30 17  18  11  12  
Cyclin A          
    <10% 25 0.0320 0.0815 0.0075 0.0353 
    ≥10% 35 21  17  12  14  
Skp2/cyclin A          
    Neither or either ≥10% 35     <0.0001 0.0004 
    Both≥10% 23     11  11  
Cyclin E          
    <10% 38 19 0.8624 13 0.3306 0.6670 11 0.7979 
    ≥10% 23 11  10    
Cks1          
    <20% 31 14 0.3443 12 0.8295 0.3281 10 0.8766 
    ≥20% 30 16  11    
p27Kip1          
    <50% 29 16 0.5155 12 0.5642 87 0.6362 0.5528 
    ≥50% 32 14  11    
Ki-67          
    <30% 39 20 0.6634 14 0.0781 0.5324 13 0.8866 
    ≥30% 20     
Mcm2          
    <50% 43 22 0.8333 13 0.0304 10 0.8771 11 0.3313 
    ≥50% 17     
Fig. 3.

Kaplan-Meier plots according to margin status (A) to predict LRFS; FNCLCC grading system [grade 1 versus 2 versus 3 (B); grade 1 versus 2 + 3 (C)] to predict DSS that showed no significant difference between grade 1 and grade 2 cases; and Skp2 expression to predict metastasis-free survival (D), DSS (E), and OS (F), respectively. In addition, the concurrent overexpression of Skp2 and cyclin A was highly predictive of DSS (G) and OS (H) for the subgroup of grade 1 and grade 2 myxofibrosarcomas.

Fig. 3.

Kaplan-Meier plots according to margin status (A) to predict LRFS; FNCLCC grading system [grade 1 versus 2 versus 3 (B); grade 1 versus 2 + 3 (C)] to predict DSS that showed no significant difference between grade 1 and grade 2 cases; and Skp2 expression to predict metastasis-free survival (D), DSS (E), and OS (F), respectively. In addition, the concurrent overexpression of Skp2 and cyclin A was highly predictive of DSS (G) and OS (H) for the subgroup of grade 1 and grade 2 myxofibrosarcomas.

Close modal

In univariate analyses, the strongest factor to predict inferior LRFS was positive margins (P = 0.0003, Fig. 3A) followed by overexpression of cyclin A (P = 0.032) whereas both the percentage of myxoid area (P = 0.0653) and Skp2 overexpression (P = 0.0585) only reached marginal statistical significance. A number of variables significantly correlated with worse MeFS, including Skp2 overexpression (P < 0.0001; Fig. 3D), remarkable tumor necrosis (P = 0.0005), high mitotic rate (≥20/10 HPFs; P = 0.0025), high grade (P = 0.0078), prominent nuclear pleomorphism (P = 0.0129), tumor size ≥6 cm (P = 0.0143), positive margins (P = 0.0295), high Mcm2 LI (P = 0.0304), and AJCC stage III (P = 0.0470). For sarcoma-related DSS, significant adverse factors were high mitotic rate (P = 0.0006), Skp2 overexpression (P = 0.0010; Fig. 3E), cyclin A overexpression (P = 0.0075), high grade (P = 0.0174; Fig. 3C), remarkable tumor necrosis (P = 0.0324), and positive margins (P = 0.0438). With respect to OS, Skp2 overexpression (P = 0.002; Fig. 3F) emerged as the strongest negative factor, followed by high mitotic rate (P = 0.006), positive margins (P = 0.0084), high grade (P = 0.02), cyclin A overexpression (P = 0.0353), and prominent nuclear pleomorphism (P = 0.0412).

Moreover, we found that co-overexpression of Skp2 and cyclin A defined a highly aggressive subset of myxofibrosarcomas that showed extremely worse DSS (P < 0.0001) and OS (P = 0.0004). The significance of poor DSS (P = 0.0006; Fig. 3G) and OS (P = 0.0093; Fig. 3H) in the Skp2-high/cyclin A-high group analysis was also robust among patients with grade 1 or grade 2 tumors. This finding might provide a very helpful prognostic value for cases at the low-grade end because FNCLCC grading scheme failed to significantly separate grade 1 from grade 2 tumors with respect to patient survival (P = 0.0908 for DSS, P = 0.0842 for OS; Fig. 3B).

Multivariate survival analyses. By multivariate analyses (Table 4), only positive margins (P = 0.0012) remained as an independent prognosticator of poor LRFS with a 3.44-fold increased risk whereas cyclin A overexpression lost statistical significance. Skp2 overexpression (P = 0.0012 for MeFS, P = 0.0056 for OS) and positive margins (P = 0.0471 for MeFS, P = 0.0173 for OS) were identified as factors independently predictive of both adverse MeFS and OS. As for sarcoma-related DSS, Skp2 overexpression (P = 0.0234), positive margins (P = 0.0152), and high mitotic rate (P = 0.0430) were independently associated with a poorer outcome. Of note, Skp2 overexpression overshadowed the prognostic influence of most clinicopathologic variables that were significant at the univariate level. Furthermore, this aberration identified patients at 13.04-fold, 7.54-fold, and 6.69-fold increased risks of worse MeFS, DSS, and OS, respectively, suggesting that Skp2 overexpression is highly representative of intrinsic biological aggressiveness of myxofibrosarcomas. On the other hand, positive margin status independently correlated with worse prognosis about all end points analyzed, implying its significant but different prognostic role in determining final outcomes.

Table 4.

Results of multivariate Cox regression analysis

FactorsLRFS
MeFS
DSS
OS
RR (95% CI)PRR (95% CI)PRR (95% CI)PRR (95% CI)P
Positive margins 3.44 (1.24-7.26) 0.0012 2.67 (1.01-7.03) 0.0471 4.55 (1.34-15.48) 0.0152 3.46 (1.24-9.61) 0.0173 
Tumor size ≥ 6 cm — — — 0.0788 — — — — 
Prominent nuclear pleomorphism — — — 0.1034 — — — 0.1700 
Necrotic area ≥ 10% — — — 0.2844 — 0.6397 — — 
Mitotic count ≥ 20/10 HPFs — — — 0.2141 4.16 (1.05—16.52) 0.0430 — 0.0900 
FNCLCC grade 2 and 3 — — — 0.3356 — 0.2141 — 0.2254 
AJCC stage III — — — 0.4347 — — — — 
Skp2 ≥ 10% — — 13.04 (2.76-61.63) 0.0012 7.54 (1.31-43.23) 0.0234 6.69 (1.74-25.65) 0.0056 
Cyclin A ≥ 10% — 0.0536 — — — 0.1632 — 0.4697 
Mcm2 ≥ 50% — — — 0.2147 — — — — 
FactorsLRFS
MeFS
DSS
OS
RR (95% CI)PRR (95% CI)PRR (95% CI)PRR (95% CI)P
Positive margins 3.44 (1.24-7.26) 0.0012 2.67 (1.01-7.03) 0.0471 4.55 (1.34-15.48) 0.0152 3.46 (1.24-9.61) 0.0173 
Tumor size ≥ 6 cm — — — 0.0788 — — — — 
Prominent nuclear pleomorphism — — — 0.1034 — — — 0.1700 
Necrotic area ≥ 10% — — — 0.2844 — 0.6397 — — 
Mitotic count ≥ 20/10 HPFs — — — 0.2141 4.16 (1.05—16.52) 0.0430 — 0.0900 
FNCLCC grade 2 and 3 — — — 0.3356 — 0.2141 — 0.2254 
AJCC stage III — — — 0.4347 — — — — 
Skp2 ≥ 10% — — 13.04 (2.76-61.63) 0.0012 7.54 (1.31-43.23) 0.0234 6.69 (1.74-25.65) 0.0056 
Cyclin A ≥ 10% — 0.0536 — — — 0.1632 — 0.4697 
Mcm2 ≥ 50% — — — 0.2147 — — — — 

Abbreviations: RR, risk ratio; 95% CI, 95% confidence interval.

RT-PCR and Western blotting assays. In six cases subjected to RT-PCR assay (Fig. 4A), the levels of Skp2 mRNA were construed as up-regulated in four cases, including three FNCLCC grade 2 and one grade 3 lesions, and as normal in two cases, including one grade 1 and one grade 2 lesion. The mean expression ratios were 4.742 versus 1.621 in the up-regulated and normal groups, respectively. These data suggested that increased transcriptional expression of Skp2 preferentially occurred in myxofibrosarcomas of higher grades. Moreover, up-regulation of Skp2 mRNA seen in four cases also translated into overexpressed Skp2 protein as shown in Western blotting assay (Fig. 4B) whereas p27Kip1 protein expression was not inversely associated with Skp2 level (Fig. 4B). The latter finding, albeit based on limited cases, seemed to corroborate our immunohistochemical analyses.

Fig. 4.

A, semiquantitative RT-PCR assay for Skp2 mRNA in three primary myxofibrosarcomas (T1, FNCLCC grade 2; T2, FNCLCC grade 2; T3, FNCLCC grade 1; N, corresponding adjacent normal tissues). The expression ratio is normalized by the housekeeping PDGB transcript and defined as up-regulated if (TSkp2 / TPBGD) / (NSkp2 / NPBGD) was ≧3. B, Western blot analyses for the same cases showed matched expression levels of Skp2 protein with respect to their corresponding mRNA transcripts whereas p27Kip1 expression was not inversely related to Skp2.

Fig. 4.

A, semiquantitative RT-PCR assay for Skp2 mRNA in three primary myxofibrosarcomas (T1, FNCLCC grade 2; T2, FNCLCC grade 2; T3, FNCLCC grade 1; N, corresponding adjacent normal tissues). The expression ratio is normalized by the housekeeping PDGB transcript and defined as up-regulated if (TSkp2 / TPBGD) / (NSkp2 / NPBGD) was ≧3. B, Western blot analyses for the same cases showed matched expression levels of Skp2 protein with respect to their corresponding mRNA transcripts whereas p27Kip1 expression was not inversely related to Skp2.

Close modal

Most authors considered that no clinicopathologic factor could reliably predict LRFS of myxofibrosarcomas despite frequent local recurrences seen in this disease entity (2, 3). In contrast, our data were in keeping with Merck et al.'s (4) observation that the adequacy of surgical margins was an independent determinant of LRFS. Huang et al. (1) reported that large size, tumor necrosis, and decreased myxoid component were adverse prognosticators for myxofibrosarcomas at the low-grade end whereas FNCLCC grading system failed to reach statistical significance. These were somewhat different from Oda et al.'s (29) series on the entire spectrum of cases, documenting that high FNCLCC grade, large size, and deep location significantly correlated with worse patient survival in univariate analyses. Nevertheless, mitotic rate was the only independent conventional variable associated with mortality (29). In agreement with Oda et al.'s finding, we also found that high mitotic activity, albeit using a different cutoff, independently increased the risk of worse DSS by 4.16-fold. However, other clinical and histologic variables lost prognostic significance in multivariate analyses. Of note in our series, positive margins also independently correlated with inferior MeFS, DSS, and OS, suggesting that improved quality of primary surgery could translate into final survival benefits. This seems to be quite unique for myxofibrosarcomas because such a strong association between margin status and final outcomes was seldom validated in primary sarcomas of other types (30, 31). Actually, the effect of microscopic residual disease on patient survival only becomes significant in locally recurrent lesions with regard to most sarcomas (32).

Oda et al.'s (29) and our analyses both indicated that tumor grading provided limited, if any, independent prognostic information in the risk-stratification of myxofibrosarcomas. In the present study, we substantiated that aberrations of proteins within the Skp2/p27Kip1–associated ubiquitin-proteasome proteolytic pathway were not uncommon. In addition, expression levels of markers in this pathway were positively related to tumor cell proliferation, as determined by the LIs of Ki-67 and Mcm2 (Table 2). Furthermore, high expression in most markers, such as Skp2, Cks1, and cyclin A, was significantly more frequently observed in a subset of cases characterized by higher grade, increasing mitotic rate, increasing tumor necrosis, or more advanced tumor stages (Table 1), thereby supporting their roles in tumor progression. As opposed to other studies, we could not verify an inverse relationship between p27Kip1 and Skp2 or Cks1 (7, 10, 13, 18). However, similar results have been recently reported for urothelial carcinomas of upper urinary tract (8), Kaposi's sarcomas (20), and a subset of malignant lymphomas (16). More intriguingly, the LI of p27Kip1 did not correlate with the grade of malignancy of myxofibrosarcomas in both Oda et al.'s study (29) and ours either. This finding seemed to be contradictory to the popular dogma about the role of p27Kip1 in cancer progression (6, 15) and could be ascribed to the following reasons. First, p27Kip1 immunolabeling can be artifactually decreased due to insufficient formalin fixation in some specimens, thereby affecting the reliability of prognostic studies wherein a loss or decrease in nuclear staining is interpreted as aberrant (33). More likely, this discordance may stem from the recent understanding that the regulatory mechanisms of p27Kip1 abundance turned out to be more complex (15, 34, 35). For instance, overexpressed Jab1 in mammalian cells can specifically interact with p27Kip1 and promote its nuclear export to accelerate its degradation (35, 36). Moreover, Skp2-independent down-regulation of p27Kip1 has recently been shown to occur at the G0-G1 transition by a novel translocation-coupled cytoplasmic Kip1 ubiquitination–promoting complex (34). Despite the limited case number, no inverse association between Skp2 and p27Kip1 was also substantiated in myxofibrosarcomas by Western blotting. This seemed to lend a support for potential Skp2-independent regulatory mechanisms of p27Kip1.

By log-rank tests, we showed that only overexpression of cyclin A, Mcm2, and Skp2 was significantly associated with a poorer clinical outcome. Cyclin A, the only cyclin with a dual role in the cell cycle, is not only required for DNA replication during S-phase progression but also active in the initiation of mitosis (37, 38). In this series, the prognostic implications of cyclin A overexpression were true for LRFS, DSS, and OS in univariate analyses but lost significance in Cox regression model. In fact, this was similar to prior studies, which could only validate the prognostic value of cyclin A at the univariate level in either heterogeneous soft tissue sarcomas (38) or myxofibrosarcomas (29) specifically.

Mcm proteins constitute a hexameric complex functioning as a helicase to unwind DNA at replication forks to initiate eukaryotic DNA synthesis (21). Mcm2, an essential regulatory subunit of this complex, has been recently found to significantly correlate with cell proliferation and prognosis in a variety of human cancers (22, 23, 39). Our observations that Mcm2 LI showed a positive correlation with the increasing mitotic rate, Ki-67 LI, and histologic grade were supportive of the pilot study of Sington et al. (40) on myxofibrosarcomas. By correlation test, they found a significant inverse association between Mcm2 expression and the time to first local recurrence (40). Nevertheless, we, by the standard log-rank analyses, could not verify the effectiveness of Mcm2 in predicting LRFS but instead proved its prognostic value with respect to MeFS. This association between high Mcm2 expression and metastatic propensity had been previously described in breast and esophageal carcinomas (23, 39).

The proteolysis of cell cycle–related proteins mediated by the ubiquitin-proteasome pathway has gained increasing importance in the understanding of the pathogenesis of human cancers (6, 10, 15, 35). Skp2 oncogenic properties have also been shown in malignancies of different lineages and, to varying degrees of significance, proved associated with histologic grade, cell proliferation, or clinical outcomes (7, 8, 10, 1214, 16, 18, 20, 36, 39). Recently, Oliveira et al. (14) have identified Skp2 overexpression as an independent negative prognosticator in a series of heterogeneous types of soft tissue sarcomas. However, to our knowledge, its prognostic value has never been specifically examined for myxofibrosarcomas. In univariate analyses, we not only confirmed its prognostic effect for the entire cohort but also recognized co-overexpression of Skp2 and cyclin A as a powerful predictor to identify the highly lethal subset from the subgroup of grade 1 and grade 2 tumors. The latter observation indicated that this synergistic interaction of cumulative abnormalities might greatly confer selective advantage on sarcoma cells, thereby translating into its superior prognostic utility. Furthermore, when expressions of Skp2, cyclin A, and Mcm2 were compared with clinicopathologically defined factors in multivariate analyses, only Skp2 remained as an independent prognostic immunomarker for MeFS, DSS, and OS in myxofibrosarcomas. Moreover, the prognostic influence of most conventional variables was explained by Skp2 overexpression, substantiating that it is highly indicative of the inherent aggressiveness of myxofibrosarcomas.

Recent in vitro models have established the role of Skp2 expression in malignant transformation because ectopic expression of Skp2 in quiescent fibroblasts contributed to deregulated activation of DNA synthesis (10, 11, 15). In transformed cells, its expression can promote Thr187-phosphorylated p27Kip1 degradation, allow the generation of cyclin A–dependent kinase activity, and induce progression of S phase (10, 11, 15). Subsequently, it has become clear that a number of other phosphorylated substrates, such as transcription factors (e.g., c-Myc), tumor suppressors (e.g., p21Waf1), and DNA replication machinery components (e.g., Cdt1), can be targeted by the F-box motif in Skp2 for ubiquitin-mediated degradation (6, 10, 15, 24). Although this expanding diversity in the targeted substrates of Skp2 has reiterated its extremely versatile oncogenic function, the underlying mechanisms responsible for Skp2 accumulation in cancer tissues are relatively less clarified. Yokoi et al. (41) reported Skp2 mRNA overexpression in 83% and gene amplification of Skp2 at chromosome 5p13 in 44% of small cell lung carcinomas, indicating that DNA amplification might represent merely one of the modes leading to Skp2 overexpression. Of great interest was a recent comparative genomic hybridization study on myxofibrosarcomas, showing very high-level amplifications frequently observed in chromosome arm 5p (25). The gains of 5p were not seen in tumors of grade 1 in that series. However, they were present in tumors of higher grades, suggesting a candidate oncogene implicated in myxofibrosarcoma progression residing in this amplicon (25). Despite the small number in cases, we confirmed increased expression of Skp2 at transcriptional level in myxofibrosarcomas of higher grade. Therefore, it seems plausible to examine if Skp2 gene is the real amplified target of myxofibrosarcomas in future investigations.

In summary, both the quality of primary surgery and intrinsic property of tumors affect the prognosis of myxofibrosarcomas. Clear margins do not only correlate with improved LRFS but also translate into final survival benefits. We clearly show that Skp2 overexpression is not only an independent adverse prognosticator but also highly representative of the inherent aggressiveness of disease. Its increased expression is also reflected at the transcription level. The co-overexpression of Skp2 and cyclin A provides a robust prognostic marker for cases at the low-grade end. The lack of an inverse correlation between p27Kip1 and Skp2 or Cks1 suggests that additional cellular events may be operating in myxofibrosarcomas, requiring further elucidation of the underlying mechanisms regulating these interacting proteins.

Grant support: National Science Council, Taiwan, grant NSC93-2320-B-182A-011 and Chang Gung Memorial Hospital grant CMRPG83019.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Prof. Jiin-Haur Chuang for kindly lending his Beecher tissue microarrayer.

1
Huang HY, Lal P, Qin J, Brennan MF, Antonescu CR. Low-grade myxofibrosarcoma: a clinicopathologic analysis of 49 cases treated at a single institution with simultaneous assessment of the efficacy of 3-tier and 4-tier grading systems.
Hum Pathol
2004
;
35
:
612
–21.
2
Mentzel T, Calonje E, Wadden C, et al. Myxofibrosarcoma. Clinicopathologic analysis of 75 cases with emphasis on the low-grade variant.
Am J Surg Pathol
1996
;
20
:
391
–405.
3
Weiss SW, Enzinger FM. Myxoid variant of malignant fibrous histiocytoma.
Cancer
1977
;
39
:
1672
–85.
4
Merck C, Angervall L, Kindblom LG, Oden A. Myxofibrosarcoma. A malignant soft tissue tumor of fibroblastic-histiocytic origin. A clinicopathologic and prognostic study of 110 cases using multivariate analysis.
Acta Pathol Microbiol Immunol Scand Suppl
1983
;
282
:
1
–40.
5
Borden EC, Baker LH, Bell RS, et al. Soft tissue sarcomas of adults: state of the translational science.
Clin Cancer Res
2003
;
9
:
1941
–56.
6
Nakayama KI, Hatakeyama S, Nakayama K. Regulation of the cell cycle at the G1-S transition by proteolysis of cyclin E and p27Kip1.
Biochem Biophys Res Commun
2001
;
282
:
853
–60.
7
Langner C, von Wasielewski R, Ratschek M, Rehak P, Zigeuner R. Biological significance of p27 and Skp2 expression in renal cell carcinoma. A systematic analysis of primary and metastatic tumour tissues using a tissue microarray technique.
Virchows Arch
2004
;
445
:
631
–6.
8
Langner C, von Wasielewski R, Ratschek M, Rehak P, Zigeuner R. Expression of p27 and its ubiquitin ligase subunit Skp2 in upper urinary tract transitional cell carcinoma.
Urology
2004
;
64
:
611
–6.
9
Oliveira AM, Nascimento AG, Okuno SH, Lloyd RV. p27(kip1) protein expression correlates with survival in myxoid and round-cell liposarcoma.
J Clin Oncol
2000
;
18
:
2888
–93.
10
Gstaiger M, Jordan R, Lim M, et al. Skp2 is oncogenic and overexpressed in human cancers.
Proc Natl Acad Sci U S A
2001
;
98
:
5043
–8.
11
Sutterluty H, Chatelain E, Marti A, et al. p45SKP2 promotes p27Kip1 degradation and induces S phase in quiescent cells.
Nat Cell Biol
1999
;
1
:
207
–14.
12
Signoretti S, Di Marcotullio L, Richardson A, et al. Oncogenic role of the ubiquitin ligase subunit Skp2 in human breast cancer.
J Clin Invest
2002
;
110
:
633
–41.
13
Yang G, Ayala G, De Marzo A, et al. Elevated Skp2 protein expression in human prostate cancer: association with loss of the cyclin-dependent kinase inhibitor p27 and PTEN and with reduced recurrence-free survival.
Clin Cancer Res
2002
;
8
:
3419
–26.
14
Oliveira AM, Okuno SH, Nascimento AG, Lloyd RV. Skp2 protein expression in soft tissue sarcomas.
J Clin Oncol
2003
;
21
:
722
–7.
15
Nakayama KI, Nakayama K. Regulation of the cell cycle by SCF-type ubiquitin ligases.
Semin Cell Dev Biol
2005
;
16
:
323
–33.
16
Lim MS, Adamson A, Lin Z, et al. Expression of Skp2, a p27(Kip1) ubiquitin ligase, in malignant lymphoma: correlation with p27(Kip1) and proliferation index.
Blood
2002
;
100
:
2950
–6.
17
Kitajima S, Kudo Y, Ogawa I, et al. Role of Cks1 overexpression in oral squamous cell carcinomas: cooperation with Skp2 in promoting p27 degradation.
Am J Pathol
2004
;
165
:
2147
–55.
18
Shapira M, Ben-Izhak O, Linn S, Futerman B, Minkov I, Hershko DD. The prognostic impact of the ubiquitin ligase subunits Skp2 and Cks1 in colorectal carcinoma.
Cancer
2005
;
103
:
1336
–46.
19
Masuda TA, Inoue H, Nishida K, et al. Cyclin-dependent kinase 1 gene expression is associated with poor prognosis in gastric carcinoma.
Clin Cancer Res
2003
;
9
:
5693
–8.
20
Penin RM, Fernandez-Figueras MT, Puig L, Rex J, Ferrandiz C, Ariza A. Over-expression of p45(SKP2) in Kaposi's sarcoma correlates with higher tumor stage and extracutaneous involvement but is not directly related to p27(KIP1) down-regulation.
Mod Pathol
2002
;
15
:
1227
–35.
21
Tachibana KE, Gonzalez MA, Coleman N. Cell-cycle-dependent regulation of DNA replication and its relevance to cancer pathology.
J Pathol
2005
;
205
:
123
–9.
22
Dudderidge TJ, Stoeber K, Loddo M, et al. Mcm2, Geminin, and KI67 define proliferative state and are prognostic markers in renal cell carcinoma.
Clin Cancer Res
2005
;
11
:
2510
–7.
23
Gonzalez MA, Pinder SE, Callagy G, et al. Minichromosome maintenance protein 2 is a strong independent prognostic marker in breast cancer.
J Clin Oncol
2003
;
21
:
4306
–13.
24
Li X, Zhao Q, Liao R, Sun P, Wu X. The SCF (Skp2) ubiquitin ligase complex interacts with the human replication licensing factor Cdt1 and regulates Cdt1 degradation.
J Biol Chem
2003
;
278
:
30854
–8.
25
Idbaih A, Coindre JM, Derre J, et al. Myxoid malignant fibrous histiocytoma and pleomorphic liposarcoma share very similar genomic imbalances.
Lab Invest
2005
;
85
:
176
–81.
26
Mentzel T, van den Berg E, Molenaar WM. Myxofibrosarcoma. In: Fletcher CDM, Unni KK, Mertens F, editors. WHO classification of tumors—pathology and genetics, tumours of soft tissue and bone. Lyon: IARC Press; 2002. p. 102–3.
27
Guillou L, Coindre JM, Bonichon F, et al. Comparative study of the National Cancer Institute and French Federation of Cancer Centers Sarcoma Group grading systems in a population of 410 adult patients with soft tissue sarcoma.
J Clin Oncol
1997
;
15
:
350
–62.
28
Soft Tissue Sarcoma. In: American Joint Committee on Cancer, editor. Cancer Staging Manual, 6th ed. Philadelphia: Lippincott-Raven; 2002. p. 221–8.
29
Oda Y, Takahira T, Kawaguchi K, et al. Altered expression of cell cycle regulators in myxofibrosarcoma, with special emphasis on their prognostic implications.
Hum Pathol
2003
;
34
:
1035
–42.
30
Trovik CS, Bauer HC, Alvegard TA, et al. Surgical margins, local recurrence and metastasis in soft tissue sarcomas: 559 surgically-treated patients from the Scandinavian Sarcoma Group Register.
Eur J Cancer
2000
;
36
:
710
–6.
31
Zagars GK, Ballo MT, Pisters PW, et al. Prognostic factors for patients with localized soft-tissue sarcoma treated with conservation surgery and radiation therapy: an analysis of 1225 patients.
Cancer
2003
;
97
:
2530
–43.
32
Gronchi A, Casali PG, Mariani L, et al. Status of surgical margins and prognosis in adult soft tissue sarcomas of the extremities: a series of patients treated at a single institution.
J Clin Oncol
2005
;
23
:
96
–104.
33
De Marzo AM, Fedor HH, Gage WR, Rubin MA. Inadequate formalin fixation decreases reliability of p27 immunohistochemical staining: probing optimal fixation time using high-density tissue microarrays.
Hum Pathol
2002
;
33
:
756
–60.
34
Kamura T, Hara T, Matsumoto M, et al. Cytoplasmic ubiquitin ligase KPC regulates proteolysis of p27(Kip1) at G1 phase.
Nat Cell Biol
2004
;
6
:
1229
–35.
35
Chamovitz DA, Segal D. JAB1/CSN5 and the COP9 signalosome. A complex situation.
EMBO Rep
2001
;
2
:
96
–101.
36
Shintani S, Li C, Mihara M, Hino S, Nakashiro K, Hamakawa H. Skp2 and Jab1 expression are associated with inverse expression of p27(KIP1) and poor prognosis in oral squamous cell carcinomas.
Oncology
2003
;
65
:
355
–62.
37
Yam CH, Fung TK, Poon RY. Cyclin A in cell cycle control and cancer.
Cell Mol Life Sci
2002
;
59
:
1317
–26.
38
Huuhtanen RL, Blomqvist CP, Bohling TO, et al. Expression of cyclin A in soft tissue sarcomas correlates with tumor aggressiveness.
Cancer Res
1999
;
59
:
2885
–90.
39
Kato H, Miyazaki T, Fukai Y, et al. A new proliferation marker, minichromosome maintenance protein 2, is associated with tumor aggressiveness in esophageal squamous cell carcinoma.
J Surg Oncol
2003
;
84
:
24
–30.
40
Sington JD, Freeman A, Morris LS, et al. Minichromosome maintenance protein in myxofibrosarcoma.
Mod Pathol
2004
;
17
:
235
–40.
41
Yokoi S, Yasui K, Saito-Ohara F, et al. A novel target gene, SKP2, within the 5p13 amplicon that is frequently detected in small cell lung cancers.
Am J Pathol
2002
;
161
:
207
–16.